Cyclic loading? Some guidance please.

[robsalv]

This is a typical one month trend of one of our catalyst filled drier vessesls.

I need to assess just how much credence I give to thermal and cyclic loading degradation concerns for this vessel.

The pink line is operating temperature - it varies between ~20DegC (normal operation) to ~200degC (regen cycle) with a temperature rate change of 25 - 30 DegC/hr. The blue line is operating pressure, which varies between ~450kPa (regen cycle) to ~3150kPa (normal operation).

The vessel is approximately 1.8m diam x 5.6m x ~36mm thick, 100% radiographed and stress relieved, with a DT=343 DegC & DP = 3999kPa. Material is A515 CS and is now approaching 37yrs old.

I've yet to have a good look in ASME VIII, but I've had a look through some other resources and this vessel seems to fall outside of the cyclic loading requirements, because it doesn't go through full cycles in either pressure or temperature, but I'm still not convinced that this vessel doesn't need to be treated with cyclic kid gloves. For the record, the original vessel design doesn't have any specific cyclic design aspects built in.


One resource recommended a cyclic design review for a cyclic deltaT of >93degC, and also recommended using a low temperature rate change to alleviate concerns - but it didn't give any indication for what that rate should be.

API579's cyclic thermal considerations were a bit more convoluted - but I suspect that at the heating rate mentioned above, the whole vessel pretty much warms up gradually and almost homogeneously so thermal issues might not come into play.


Any helpful hints, thoughts or guidance would be greatly appreciated.


Thanks in advance.

 

[dcasto]

This is a very typical regeneration system used with mole sieves. I've seen vessels like like well over 40 years old. The heat and cool rates are so slow, its almost like getting a heat treatment every few days. Based on the age, are there vessels internally insulated, at majority were in the time frame you set.

I was also wondering why the beds were heated and cooled twice in the first two regenerations?

 

[davefitz]

A simplified set of rules for determining the amount of fatigue damage in cyclindrical vessels can be obtained from the german boiler code TRD 301 annex 1 or the PED newer code EN 12952-3.

Basically, the thermal stress will be proportional to teh square of the wall thickness times the linear ramp rate in temperature. The max allowable alternating stress range is related to the number of cycles prior to initiation of the first visible crack, However, if the highest stress occurs at a weld , then a pre-exisitng weld crack is postulated to exist, and fracture mechanics -related fatuigue curves are then used.

 

[robsalv]

Thanks for the replies.

Dcasto, they aren't internally insulated. I'm not aware of why the temperature profile doubled up like it did... but it means several more cycles a year to consider.

Davefitz, I'll keep that reference in mind if I can't find a more close at hand answer to whether I need to consider this vessel with full cyclic considerations.

Just a bit more info, Wet Fluro mag has found some indications in the past, but none were deemed fatigue like in nature. Some LOF's broke through in a couple of nozzle welds and some indications were picked up on the catalyst bed support ring attachment welds on the ring side.

 

[robsalv]

Well, I finally found the fatigue evaluation parts of ASMEVIII, in Div2, clause AD-160. ASME VIII is not our normal code in Australia, even though the vessel in question was an ASMEVIII vessel.

Interestingly, the fatigue evaluation section is virtually word for word identical with the similar section of the Australian pressure vessel code, AS1210.

The "20% variation of design pressure" clause is the one that snags this vessel. According to the current version of the code, if there are more than 1000 cycles expected, then a fatigue analysis shall be performed. This vessel is clearly over that threshold... but having a longer in detail look through the screening parts of API579, appendix B and F, it looks like this vessel hasn't tripped API579's fatigue screening threshold yet.

I think I'm convinced the vessel should be investigated for cyclic issues even though it originally wasn't designed for it... probably more an artefact of missing requirements in 1960 rather than by design.

Happy to take more comments.

Cheers